Constitutional Commitment to Social Security and Welfare Policy

Constitutional Commitment to Social Security and Welfare Policy1

Avi Ben-Bassat∗

Momi Dahan**

February 2006

ABSTRACT

In this paper we explore whether the constitutional text has any practical meaning for

welfare policy. To examine the empirical importance of the constitution, we first

constructed for 68 countries an index of constitutional commitment to social security

in five areas: Old Age, disability and survivors (OASDI), Unemployment, Sickness,

Work Injury and Income Support. We find that the extent and coverage of social

security laws is not sensitive to the degree of constitutional commitment to social

security.

Key words: Welfare Policy, Social Security, Constitution, Legal Origins

1 We wish to thank Margit Cohen and Yishay Mishor for their helpful comments. We also thank Liza Teper for her research assistance. We thank the Minerva Center for Human Rights for financial support. ∗ Department of Economics, Hebrew University of Jerusalem, Israel 91905. **School of Public Policy, Hebrew University of Jerusalem, Israel 91905, [email protected]

2

Introduction The history of constitutional human rights is stratified into three generations. Civic

rights such as the freedom of expression are perceived as the first generation, political

rights such as the right to vote comprise the second generation, and social rights such

as the right to live in dignity are the third generation. Most countries include the first

two generations of rights in their constitution, but they differ significantly in the scope

and degree of commitment to social rights. In some countries there is not even a

mention of social rights in their constitutions.

There is extensive literature on the consequences of electoral rules and forms of

government as outlined in the constitution on economic policy outcomes. For

example, Persson and Tabellini (2004) found that constitutional rules have a

significant effect on key fiscal policy and corruption.2 The implication is that the

constitutional text has more than a declarative role in this regard. But does the

constitution have any practical meaning when it comes to social rights?

In a companion paper (Ben-Bassat and Dahan, 2003) we have constructed indices to

assess the degree of constitutional commitment to five social rights (Social Security,

Education, Health, Housing and Workers’ Rights). While we did not trace a robust

and significant effect of constitutional commitment to education and health on

education and health expenditures, we did find a positive relationship between the

degree of constitutional commitment to social security and transfer payments (as a

share of GDP). This finding is consistent with the claim that the constitution is more

than a piece of paper in the sense that the constitutional text might be important for

public policy.

The positive association between constitutional commitment to social security and

transfer payments motivated us to go a step further to examine the transmission

mechanism. If the constitution does have an effect on welfare policy, this should

emerge in the connection between the degree of commitment to social security in the

constitution and the extent and coverage of social security laws. The importance of

the constitutional text for policy is in question if it does not survive this test.

2 See Persson and Tabellini (2003) book for an excellent survey of the empirical evidence.

3

The goal of this paper is to explore whether the constitutional text regarding social

security plays a role in shaping social security policy. Social security is one of the

main pillars of the modern welfare state. In fact, social security expenditures account

for a large chunk of government outlays in a typical developed country. The European

average share of social security expenditure is 24.8% of GDP while the world average

is 14.5% (ILO, 2000).

Most social security programs are formulated as entitlements, and so it helps to

quantify the extent and coverage of social security programs in a relatively

meaningful way. Naturally, the constitutional commitment to a particular right must

be examined in terms of the associated social security law. This makes room for a

range of effects of constitutional commitment on various components of social

security. For example, the constitutional commitment to Old Age, compared to

Unemployment, may have a different effect on social security policy. This type of

detailed examination provides a more stringent test of the importance of constitutional

text on social policy.

In the next section we present the criteria used to translate the constitutional text into

quantitative indices that reflect the constitutional commitment to social security. In

Section 3 we examine whether our indices are correlated with legal origins and

religious beliefs. In Section 4 we look at the effect of constitutional text on the extent

and coverage of social security laws. Section 5 concludes the paper.

2. Indices of Constitutional Commitment to Social Security In this section we construct constitutional indices of social security rights according to

the constitutional text, and ignoring court interpretations. There is wide variability in

constitutional social security rights, ranging from the U.S. and Australia, where social

security rights are absent, to Switzerland and Brazil, which have a high constitutional

commitment to social security.

Our paper relates to 64 countries with a written constitution and four which have a

legal document with a higher status than regular law. In Canada, New Zealand and

4

Israel there are basic laws which have a legal status similar to a constitution. England

does not have basic laws, but it does have a Human Rights Act that has higher status

than regular law.3 Our two main sources are the English translation of the constitution

in the ICL and Confinder web sites.4

A constitutional social right is defined here as one that grants a personal entitlement to

monetary transfers (including social insurance). That right may affect permanent

income and welfare. Those social rights provide a social safety net and would seem to

have a positive impact on income equality, at least in the short run.

There are five groups of social security rights in a constitution, and each of them may

contain one or more features. The five rights are the following:

1. The three most basic features of social security as reflected in

textbooks in Public Economics 5 are: pension, disability and survivors

(OASDI)

2. The right to live in dignity, usually expressed by income support to the

poor

3. Unemployment insurance

4. Sickness insurance

5. Work injury insurance

There is considerable variance between countries as regards the degree of

constitutional commitment to social security rights, ranging from concrete policy

action in some countries to a general statement reflecting a vague commitment in

others. We rank the degree of constitutional commitment on a scale from 0 to 3, with

a rank of 0 denoting a right that is absent from the constitution.

A rank of 1 is given if the constitution includes a general statement with regard to a

particular social security right. In that case it is clear that it is possible to introduce a

3 From Section 3 in the English Human Rights Act it can be inferred that ordinary laws are subject to the Human Rights Act. Any law should be examined in the light of the Human Rights Law. In case the suggested law is in contradiction to the Human Rights Act, the law may still be passed, provided the parliament is aware of this. 4 International Constitutional Law http://www.uni-wuerzburg.de/law 5 "This [OASDI] is usually referred to as social security and is intended to provide a basic standard of living to the old aged, the disabled, and their survivors." (Stigliz, 2000: page 353).

5

law concerning that right. A rank of 2 is given if the constitution guarantees a minimal

level with respect to that right such as “a minimum standard of living,” or “a life of

dignity,” in the case of income support and “adequate insurance” in the case of

unemployment. A rank of 3 is given if the constitution has a high degree of

commitment and concreteness. For example, a commitment “to maintain the real

value” of social security benefits justifies a rank of 3.

We use the two most common constitutional social security rights – the right to

pension in old age and the right to unemployment insurance – to illustrate the ranking

process.

The right to old age, survivors and disability (OASDI)

Each of the three features constituting the right to OASDI was ranked as shown in

Table 1. The overall rank is a simple average across all three features. The most

popular rights are old age pension (41 countries), and assistance to the disabled (39

countries), while support to survivors is less frequent (28 countries).

Table 1: Criteria for ranking the constitutional commitment to old age, survivors

and disability (OASDI)

rank

The right is absent from the constitution 0

General statement

The state “guarantees” or “promotes” social security insurance, or

“every person is entitled to social security”

1

Weak commitment

Old age, disabled and survivors are entitled to a “minimum standard

of living,” “basic income,” “adequate income” or “to live in

dignity”

2

Strong commitment

In addition to the guarantee of “adequate income,” the constitution

specifies the elements of what is adequate income in terms of food,

housing etc., or a periodical adjustment mechanism such as COLA.

3

6

We use the constitution of Switzerland as an example of a country that receives a rank

of 3 as regards OASDI rights. To quote the Swiss constitution: “The confederation

shall take measures for adequate social security for the elderly, survivors, and

disabled persons. The insurance shall be mandatory. The pensions must cover basic

living expenses appropriately. The maximal pension shall not exceed twice the

minimal pension. The pension shall at least be adapted to the development of prices .”

The key elements that are responsible for its high rank are: the insurance is

mandatory, there are limits on minimal amounts and the pension is adjusted to prices.

The constitution of Italy guarantees support just to two groups usually included in

OSDAI – the elderly and the disabled, as may be seen from the following quote:

“Workers shall be entitled to adequate insurance for their needs in case of accident,

illness, disability, old age, and involuntary unemployment.” The constitution assures

that workers will receive adequate insurance, but it does not mention any further

details such as periodical adjustment. Because of its weak commitment it gets a rank

of 2 for each group covered, but since only two out of three potential groups are

covered the overall rank is 1.333.

India’s constitution contains a general statement concerning the two features of the

right to OASDI; each of them earns a rank of 1, as is indicated by the following

quote: “The state shall, within the limits of its economic capacity and development,

make effective provision for securing the right to public assistance in cases of

unemployment, old age, sickness and disablement.”

The right to unemployment insurance

The commitment to unemployment insurance is usually formulated along the same

lines of other social security rights, such as old age pension and sickness insurance. In

many constitutions they are even formulated as part of the same article, as shown for

Italy and India in the previous section. Therefore, the unemployment insurance

commitment was ranked by the same scale presented in Table 1.

Brazil is an excellent example of a country that is strongly committed to

unemployment insurance (rank: 3), as is indicated by the following quote: “The

social security plans shall, upon contribution, pursuant to the law, provide protection

7

of workers who are involuntarily unemployed. Adjustment of the benefits is ensured

so as to permanently maintain their real value according to criteria defined in the

law.”

The commitment of Hungary is weaker (rank: 2), as expressed in the following quote:

“Citizens of the Republic of Hungary have the right to social security; they are

entitled to the support required to live in old age, and in the case of sickness,

disability, being widowed or orphaned and in the case of unemployment through no

fault of their own.” Hungary grants entitled citizens “the support required to live,” but

do not provide any extra element, such as a periodical price adjustment.

A rank of 1 is given if the constitution includes a general statement with regard to

unemployment insurance. A typical example appears in the constitution of Egypt –

“All citizens have the right to pensions in cases of incapacity, unemployment, and old

age, in accordance with the law.” It does not include any characteristic beyond a

general statement and it is even more restricted in the sense that the translation of this

right should be consistent with the (regular) law.

The constitutional commitment to all social security rights were ranked according to

these criteria, and the data set for each characteristic are presented in Table 1.

3. Commitment to social security, traditions and religious beliefs

The constitutional commitment to social security may be regarded as the core of the

modern welfare state. The right to social security appears in the constitution of 47

countries with different levels of commitment. In half the countries, the rank value is

less than 1 and the average is 0.56, reflecting a relatively low level of constitutional

commitment. We found a high level of constitutional commitment in four countries:

Brazil, Finland, Portugal and Switzerland.

The constructed index of constitutional commitment to social security is composed of

five features: Old Age (including survivors and disability), Unemployment, Sickness,

8

Work Injury and Income Support. The most prevalent social security right in the

constitutions is Old Age, which appears in 43 countries, while Work Injury insurance

appears only in 11 countries.

3.1 CCSS and legal origins At first glance it is hard to find common economic, cultural or other characteristics

among countries that share a similar degree of constitutional commitment to social

security (Table 1). For example, the Scandinavian countries are spread all over the

scale. Finland is closer to Latin-American countries, Sweden has the same

constitutional commitment as South Korea, while Norway belongs to a group of

countries in which social security is absent from the constitution.

All of the countries in our sample were classified by legal origins, in accordance with

the groups suggested by Reynolds and Flores (1989). The two main legal traditions

are English common law and French civil law, derived from Roman law. We

followed a series of studies showing the importance of legal origins for economic

performance [La-Porta et al (1997, 1998, 1999), Glaeser and Shliefer (2002) and

Botero et al (2004)].

The concept underlying the Common Law tradition is to protect citizens from the

power of government. It began to develop in the 17th century, with the empowerment

of the parliament and aristocracy at the expense of the monarchy, and concomitant

greater constraints on the power of the king (Finer, 1997). In contrast, the civil law

tradition, especially after codification in the 19th century, gives the government more

power to run the life of its citizens (Finer, 1997). There are three groups of countries

that follow the civil law tradition—French, Scandinavian and German.

In Eastern Europe the legal tradition is relatively new and its roots are in the former

socialist pattern of the Soviet Union, which is far more centralized than civil law.

Each country in our sample is classified into one of the five groups according to its

legal tradition: English common law, French civil law, German, Scandinavian and

socialist.

9

It is striking that seven out of the top ten countries ranked according to our summary

index of constitutional commitment to social security are French civil law countries:

Portugal, Italy and five Latin American countries (Table 2). This finding, with regard

to Latin American countries, is surprising in light of the high level of income

inequality in those countries.

In contrast, no common law country is part of the top ten. According to our

constitutional commitment index, there are 21 countries where all five social security

rights are absent. Thirteen of those are common law countries, while only four are

French civil law countries.

In Table 3 we test the hypothesis that the constitutional commitment to social security

rights is related to legal origins, controlling for level of development and propensity to

democracy. All equations are estimated with OLS, where each social security right

serves as a dependent variable at one time. We find that countries that are classified as

French civil law have a much higher constitutional commitment to social security than

do common law countries (the omitted variable), after controlling for GDP per capita

and democracy. This is a highly significant result.

This finding repeats itself for all five social security rights separately (Table 3). The

coefficient reflects quite a large effect: The coefficient for the summary index of

constitutional commitment to social security is around 0.61, where the sample mean

of our index is 0.56. Countries belonging to the socialist tradition, Scandinavian or

German legal origins do not express different constitutional commitment to social

security as compared to common law countries. At first glance, it is somewhat

surprising to find that French civil law countries have on average a higher

constitutional commitment to social security than post-socialist countries. Most

socialist countries are in transition to a market economy, however, and some of them

have rewritten their constitutions in recent years.

Given the fact that the top ten (or even twenty) countries are disproportionably

populated by Latin American countries, it is natural to examine the results’ sensitivity

to the inclusion of a dummy variable for Latin American countries. We found that in

general the results are similar (Table 4). However, the introduction of a Latin

10

American dummy variable leads to a lower coefficient for French civil law countries,

without a change in ordering: countries with a socialist tradition, Scandinavian or

German legal origins are not statistically different from common law countries.

3.2 CCSS and religious beliefs

It is natural to examine whether the constitutional text is related to the religious

beliefs of each society.6 We use the shares of population that have Protestant,

Catholic, Muslim and other beliefs as explanatory variables controlling for GDP per

capita. Table 5 presents OLS regressions for each of the five constitutional

commitments to social security features in addition to a summary index of social

security.

The picture that emerges here is much less clear than that regarding legal origins. In

general we find that countries which have a higher share of population with Catholic

beliefs tend to have higher constitutional commitment to social security as compared

to Protestant (and other beliefs) countries.

However, that effect is statistically significant only regarding sickness and work

injury. The effect of Catholic beliefs for the other three social security rights is

positive but insignificant. The Catholic coefficient is not significant at the standard

level in the regression of our summary index of constitutional commitment to social

security.

The quantitative impact of Catholic beliefs is quite large, but only for those

coefficients that are significant. In addition, the Catholic effect does not survive the

inclusion of a dummy variable for Latin American countries (Table 6). In the case of

the constitutional commitment to work injury, the Catholic coefficient becomes

insignificant, and in the case of sickness the quantitative impact implied by the

estimated coefficient is much less.

6 Note that there is some overlapping between legal origins and religious beliefs, and in particular between Catholic and French civil law.

11

4. Social security in the constitution and in practice

The commitment to social security in the constitution might reflect social preferences

but does not necessarily translate into government policy. Naturally, the constitution

does not imply concrete policy actions, delineating a path for policy makers. There

may, therefore, be a weak or even no relationship between constitutional commitment

and welfare policy.

This paper examines the empirical correlation between our indices of constitutional

commitment to social security and the generosity implied in five features of social

security laws. If the constitutional text with respect to social insurance has any

practical meaning it should appear in the extent and coverage of social security laws.

We should expect to have a positive correlation between welfare policy indicators and

our index of constitutional commitment to social security if the constitutional text has

any practical meaning. The danger of reverse causality arises in most studies using

regression analysis. However, the risk of reverse causality is less likely in our case to

the extent that constitutions reflect beliefs and values. While government policy may

lead to changes in the constitution, this is rare. Values and beliefs may change over

time, but this is a slow process that does not necessarily have an immediate impact on

the constitution.

Alternatively, no relations between the constitutional text and welfare policy may

reflect the fact that the constitution has a solely declarative role. Various policies

might be consistent with the same constitutional commitment to social security.

Unlike civic and political rights, social security rights might be less binding because

of the qualitative nature of these rights.

No empirical relations could be also the result of the fact that public policy is shaped

by social preferences which are not yet in the constitution, and may never be. In this

case, we do not expect to find any relation between welfare policy and social security

rights (which by construction equal zero), unless there is a systematic bias in the sense

12

that countries that are more welfare-state oriented tend to omit social security rights

from their constitution.

The last hypothesis—a somewhat cynical one—is that the constitutional text

embodies the very opposite of the real beliefs that dictate welfare policy. In such

regimes the constitution is merely of propaganda value. In this case we would expect

to find a negative relation between constitutional commitment to social security and

the extent of social insurance coverage.

Before turning to regression analysis, Table 7 presents a simple matrix for four

different features of social security: Old age, Unemployment, Work Injury and

Sickness. The columns describe whether a particular feature exists in the constitution

(regardless of its degree of commitment), and the rows represent whether that same

feature appears in the regular laws as they are documented in Social Security

Throughout the World 2004.

Panels A and B of Table 7 show that Old Age (including survivors and disability) and

Work Injury laws exist in all countries. Those countries that have not mentioned Old

Age (25 countries in our sample) and Work Injury (57 countries in our sample) in

their constitution still provide these social insurance features. This simple matrix

implies that the lack of constitutional commitment to a certain social security feature

is not translated to an absence of that particular feature from social security law.

The other off-diagonal cells in four panels of Table 7 are even more interesting. They

answer the question of whether there are countries that have expressed commitment to

particular features of social security and yet do not provide coverage at all. We find 6

such countries in the case of Unemployment and two countries in the case of

Sickness. However, it is hard to draw conclusions from this finding due to data

limitations. Unfortunately, the Social Security Throughout the World publication

pooled together countries that do not have a particular feature of social security with

countries for which the data is unavailable.

Our focus here is to examine the effect of constitutional commitment to social

security on welfare policy, controlling for a list of control variables that is standard in

13

this literature.7 Our control variables are GDP per capita, the propensity to democracy

and the share of the population over age 65.

The data for GDP per capita is an average for 1990–1999, taken from the Penn World

Tables for all countries except Taiwan. Income per capita serves as an indicator of the

level of economic development, which may influence social preferences for public

consumption (versus private consumption), as well as a more developed tax-collection

system (‘Wagner’s law’).

A large share of social security expenditure goes to the population over the age of 65.

Hence, the average share of the population aged above 65 for 1990 –1999 is one of

the control variables (taken from World Bank Data).

Regression analysis

In this paper we provide two ways to examine the relationship between constitutional

commitment to social security and welfare policy. In general, expenditures such as

transfer payments are a complex combination of the extent of social security coverage

as determined by social security laws and “exogenous variables” such as take-up rate,

demographic composition and aggregate unemployment rate (in the case of

unemployment expenditures). We first use the following econometric model to test

the effect of constitutional commitment to social security:

(i) ε+++= CCSSaYaaTR 210 ,

where TR denotes government expenditures on actual transfer payments, Y is a vector

of various control variables such as demographic composition, and CCSS stands for

constitutional commitment to social security.

The coefficient of CCSS captures the pure effect of social security laws to the extent

that we control for exogenous variables, which is not always possible. In addition, this

7 For example, Rodrik (1998), Taveres and Wacziarg (2001) and Mulligan, Gil and Sala-i-Martin

14

type of econometric model assumes that transfer payments are a linear combination of

social security laws and exogenous variables.

Alternatively, we use a somewhat more direct way to estimate the effect of

constitutional commitment to social security on the extent and coverage of social

security laws:

(ii) uCCSSZLAWS 210 +β+β+β= ,

where LAWS represents an index of the generosity of social security laws, Z is a

vector of various control variables, and CCSS is as before.

In an estimated model of type (ii) the empirical challenge is to construct a meaningful

index that captures the extent and coverage of social security laws. Note that the

generosity of social security laws may be related to the size of the eligible population.

Social entitlements may be more generous with a larger eligible population because of

political considerations through a median voter effect, or less generous due to budget

considerations.

Table 8 presents several regressions for various indices of the generosity of social

security. In this first column we replicate the Ben-Bassat and Dahan (2003)

regression, where a summary index of constitutional commitment to social security

was found to be with a positive coefficient in the share of transfer payments in GDP.

The estimated coefficient reflects a substantial quantitative effect. An increase of one

standard deviation in the constitutional commitment to social security would induce a

rise of 1.7 percentage points in the share of transfers in GDP if this correlation reflects

a causal relationship.

The effect of GDP per capita on transfer payments is positive (as expected by

Wagner’s law), but it is significant only when the share of the population over age 65

is not included. However, the positive effect of the elderly population on transfer

payments is not sensitive to the inclusion of other variables, such as democracy. The

(2002).

15

insignificant effect of democracy on social spending was found also in Mulligan, Gil

and Sala-i-Martin (2002).

Nevertheless, transfer payments as a proxy for the extent and coverage of social

security suffers from two main limitations. First, transfer payments include social

security expenditures in addition to other government transfers. Second, the issue of

data availability makes it hard to control for all the relevant factors. It is important to

the extent that the error term in a model of type (i) described above is correlated with

the omitted variables. Third, the international institutions that collect and publish

budgetary data are always careful to add footnotes regarding the limitations of

comparing fiscal data across countries. In fact, we found large discrepancies between

various sources in the data on social security expenditures or transfer payments.

In column 2 (Table 8) we run a regression where the dependent variable is the

contribution rate for all social security programs (taken from Social Security

Throughout the World 2004). The contribution rate as an indicator for the extent and

coverage of social security programs is more comparable across countries than

transfer payments.

Nonetheless, the limitation of that proxy is clear: social security benefits may be more

generous than what is implied by the contribution rate in countries that have

experienced long and substantial actuarial deficits. However, Mulligan and Sala-i-

Martin (2004) report on a relatively large correlation between the amount of revenue

collected by payroll taxes and social security expenditures. We find, however, that the

coefficient of constitutional commitment to social security becomes insignificant once

we replace transfer payments with the contribution rate.

In a recent paper, Botero et al (2004) constructed an index of social security coverage

for 85 countries. We use their summary index of social security coverage instead of

transfer payments as a proxy for social security coverage (52 countries exist both in

their and our samples). Once again, the coefficient of constitutional commitment to

social security turns out to be insignificant as in the case with the contribution rate.

16

The summary index of social security coverage by Botero et al (2004) is built on three

different indices: Old Age (including survivors and disability), Unemployment and

Sickness. It allows us to test our hypothesis more directly by regressing a particular

constitutional commitment on its respective social security feature. As before, we find

no empirical relation between constitutional commitment to social security and social

security in practice in all three features separately: Old Age, Unemployment and

Sickness (Table 9). In table 10 we go further to test the effect of constitutional

commitment to Unemployment separately on four components of the extent and

coverage of unemployment law as calculated by Botero et al (2004). The coefficients

of constitutional commitment to Unemployment are insignificant in all four

regressions.

5. Conclusion

In this paper we have constructed for 68 countries five indices of constitutional

commitments to social security rights: Old Age (including Survivors and Disability),

Unemployment, Sickness, Work Injury and Income Support. The commitment to Old

Age appears in the constitution of 43 countries while only 11 countries have a

constitutional commitment to Work Injury.

21 countries preferred not to include any social security right, but nevertheless in

practice they do provide social security. In general, in countries that have social

security rights in their constitutions the degree of commitment is relatively low; some

countries lower it further by adding that those rights should be in accordance with the

law. This may reflect the fear that constitutional commitment might undermine the

flexibility of economic policy to adjust to environment changes such as

macroeconomic fluctuations or preferences toward government involvement.

We find that countries that share the tradition of French civil law generally have a

higher constitutional commitment to social security than countries adhering to the

English common law tradition. We have not found a clear and significant difference

between German, Scandinavian or Socialist traditions and English common law

tradition.

17

In a previous paper (Ben-Bassat and Dahan, 2003) we found a significant correlation

between constitutional commitment to social security and transfer payments. If the

constitution does have an effect on welfare policy it must show up in the connection

between the degree of commitment to social security in the constitution and the extent

and coverage of social security laws. In this paper we found no support for the effect

of constitutional commitment to social security on social policy when using various

measures of social security coverage in regular laws instead of transfer payments as

an indicator for social policy.

This finding may reflect the nature of constitutional social rights. The practical

translation of constitutional civic rights such as the right to vote is clear. In contrast,

the translation of social rights such as unemployment insurance may be expressed by

various level of generosity. For example, the replacement rate and the period of

entitlement may be different even for the same degree of constitutional commitment

to unemployment insurance.

18

Table 1: Indices of constitutional commitment to social security

Summary index of Social

Security Income Support Unemployment Sickness

Work Injury

Old Age, Survivors

and Disability

Albania 0.43 1 0 0 0 0.66 Argentina 0.43 0 0 0 0 1.00 Australia 0.00 0 0 0 0 0.00 Austria 0.00 0 0 0 0 0.00 Bahrain 0.71 0 1 1 0 1.00 Belgium 0.43 0 0 0 0 1.00 Bolivia 0.86 0 1 1 1 1.00 Brazil 3.00 3 3 3 3 3.00 Bulgaria 0.43 0 1 0 0 0.66 Cameroon 0.00 0 0 0 0 0.00 Canada 0.00 0 0 0 0 0.00 Chile 0.43 1 0 0 0 0.66 China 0.86 2 0 0 0 1.33 Colombia 0.43 0 0 0 0 1.00 Cyprus 0.86 0 0 0 0 2.00 Czech Republic 0.00 0 0 0 0 0.00 Denmark 0.14 1 0 0 0 0.00 Dominican Republic 1.57 3 2 2 0 1.33 Ecuador 1.29 2 0 1 1 1.66 Egypt 0.43 0 1 0 0 0.66 El Salvador 0.43 0 0 0 0 1.00 Fiji 0.00 0 0 0 0 0.00 Finland 2.14 3 3 0 0 3.00 France 0.43 2 0 0 0 0.33 Germany 0.00 0 0 0 0 0.00 Greece 0.00 0 0 0 0 0.00 Hungary 1.43 0 2 2 0 2.00 Iceland 0.14 1 0 0 0 0.00 India 0.57 1 1 1 0 0.66 Indonesia 1.00 1 0 0 0 2.00 Iran 0.71 0 1 0 1 1.00 Ireland 0.43 0 0 0 0 1.00 Israel 0.00 0 0 0 0 0.00 Italy 1.71 2 2 2 2 1.33 Japan 0.14 1 0 0 0 0.00 Jordan 0.00 0 0 0 0 0.00 Kenya 0.00 0 0 0 0 0.00 Malta 0.86 1 1 1 1 0.66 Mexico 0.86 0 1 1 1 1.00 Nepal 0.43 0 0 0 0 1.00 Netherlands 0.14 1 0 0 0 0.00 New Zealand 0.00 0 0 0 0 0.00 Nicaragua 1.86 3 2 0 2 2.00 Norway 0.00 0 0 0 0 0.00 Panama 1.14 0 2 1 1 1.33 Paraguay 0.43 1 0 0 0 0.66 Philippines 0.00 0 0 0 0 0.00 Poland 0.57 1 0 1 0 0.66 Portugal 2.00 2 2 2 0 2.66 Romania 0.29 0 1 0 0 0.33 Sierra Leone 0.29 0 0 0 0 0.66 Singapore 0.00 0 0 0 0 0.00 South Africa 0.86 3 0 0 0 1.00 South Korea 0.43 0 0 0 0 1.00 Spain 1.00 2 2 0 0 1.00 Sri Lanka 0.00 0 0 0 0 0.00 Sweden 0.43 0 0 0 0 1.00

19

Switzerland 2.14 1 3 1 1 3.00 Syria 0.57 0 0 1 0 1.00 Taiwan 0.43 0 0 0 0 0.66 Thailand 0.00 0 0 0 0 0.00 Trinidad 0.00 0 0 0 0 0.00 Tunisia 0.43 0 0 0 0 1.00 Turkey 0.29 0 0 0 0 1.00 United Kingdom 0.00 0 0 0 0 0.00 United states 0.00 0 0 0 0 0.00 Uruguay 1.71 0 2 2 2 2.00 Zambia 0.00 0 0 0 0 0.00 Mean 0.56 0.57 0.5 0.33 0.23 0.78 Standard deviation 0.65 0.92 0.86 0.67 0.59 0.81 No. of countries with positive value 47 23 20 16 11 43 Mean for countries with positive value only 1.70 1.70 1.44 1.46 1.23

20

Table 2: Constitutional commitment to social security and legal origins (Countries are ranked by a summary index of constitutional commitment to social security)

Rank Country Legal

Origins

Summary Index of Social

Security

Rank Country Legal

Origins


Security 1 Brazil F 3.00 35 Paraguay F 0.43 2 Finland SD 2.14 36 Chile F 0.43 3 Switzerland G 2.14 37 France F 0.43 4 Portugal F 2.00 38 Colombia F 0.43 5 Nicaragua F 1.86 39 Sweden SD 0.43 6 Uruguay F 1.71 40 Tunisia E 0.43 7 Italy F 1.71 41 Turkey F 0.29

8 Dominican Republic F 1.57

42 Sierra Leone E 0.29

9 Hungary S 1.43 43 Romania S 0.29 10 Ecuador F 1.29 44 Iceland SD 0.14 11 Panama F 1.14 45 Denmark SD 0.14 12 Indonesia F 1.00 46 Netherlands F 0.14 13 Spain F 1.00 47 Japan G 0.14 14 Bolivia F 0.86 48 Austria G 0.00 15 South Africa E 0.86 49 Australia E 0.00 16 Malta F 0.86 50 United states E 0.00

17 Mexico F 0.86

51United Kingdom E 0.00

18 China S 0.86 52 Germany G 0.00 19 Cyprus E 0.86 53 Zambia E 0.00 20 Iran F 0.71 54 Trinidad E 0.00 21 Bahrain E 0.71 55 Greece F 0.00 22 India E 0.57 56 Jordan F 0.00 23 Syria F 0.57 57 Israel E 0.00 24 Poland S 0.57 58 Norway SD 0.00 25 Ireland E 0.43 59 New Zealand E 0.00 26 El Salvador F 0.43 60 Singapore E 0.00 27 Albania S 0.43 61 Sri Lanka E 0.00 28 Argentina F 0.43 62 Fiji E 0.00 29 Bulgaria S 0.43 63 Philippines F 0.00

30 Belgium F 0.43

64Czech Republic S 0.00

31 South Korea G 0.43 65 Cameroon F 0.00 32 Taiwan G 0.43 66 Canada E 0.00 33 Egypt F 0.43 67 Kenya E 0.00 34 Nepal E 0.43 68 Thailand E 0.00

21

Table 3: Legal origins and constitutional commitment to social security

(the dependent variable- an index of constitutional commitment to social security)

(6)

(5) (4) (3) (2) (1)


Security

Work Injury

Average of Old Age, Disability and Survivors SicknessUnemployment

Income Support

0.543 0.617 0.813 0.107 0.047 0.808 Constant

(0.580)

(0.740) (0.680) (0.110) (0.040) (0.601)

-0.058 -0.101 -0.051 -0.031 -0.019 -0.114 Log GDP Per-Capita

(-0.490) (-0.950) (-0.340) (-0.250) (-0.120) (-0.650)

0.271 0.407 0.055 0.385 0.328 0.567 Democracy

(0.840) (1.400) (0.130) (1.130) (0.750) (1.200)

0.612ª 0.521ª 0.648ª 0.497ª 0.669ª 0.643ª French

(3.390) (3.200) (2.800) (2.610) (2.710) (2.264)

0.334 -0.030 0.403 0.313 0.460 0.342 Socialist

(1.210) (-0.120) (1.140) (1.070) (1.220) (0.850)

0.313 0.199 0.447 0.028 0.432 0.176 German

(0.950) (0.670) (1.060) (0.080) (0.960) (0.370)

0.336 -0.007 0.448 -0.182 0.422 0.765 Scandinavian

(1.01) (-0.02) (1.050) (-0.520) (0.930) (1.570)

0.084 0.117 0.029 0.059 0.032 0.037 Adj. R2

67 67 67 67 67 67 Number of Observations

The t statistics are reported in the parentheses. a. Significance at 1%

22

Table 4: Legal origins and constitutional commitment to social security With Latin American dummy (the dependent variable- an index of constitutional commitment to social security)

(6)

(5) (4) (3) (2) (1)

Constitutional commitment to: Summary Index of Social

Security

Work Injury Sickness

Old Age, Survivors and

Disability UnemploymentIncome Support

0.060 -0.154 -0.246 0.217 -0.335 0.783 Constant

(0.060) (-0.200) (-0.240) (0.180) (-0.260) (0.550)

0.010 0.008 0.019 0.033 0.034 -0.110 Log GDP Per-Capita

(0.090) (0.080) (0.150) (0.210) (0.200) (-0.600)

0.093 0.123 0.255 -0.164 0.187 0.557 Democracy

(0.280) (0.450) (0.720) (-0.390) (0.410) (1.130)

0.430b 0.229 0.363c 0.422c 0.524c 0.634b French

(2.140) (1.360) (1.690) (1.640) (1.870) (2.100)

0.351 -0.003 0.326 0.423 0.473 0.343 Socialist

(1.300) (-0.020) (1.120) (1.220) (1.250) (0.840)

0.285 0.154 0.008 0.412 0.410 0.175 German

(0.880) (0.570) (0.020) (0.990) (0.910) (0.360)

0.310 -0.050 -0.201 0.415 0.401 0.764 Scandinavian

(0.950) (-0.180) (-0.580) (0.990) (0.880) (1.560)

0.480b 0.766ª 0.351 0.592c 0.380 0.025 Latin America

(1.930) (3.680) (1.320) (1.860) (1.100) (0.070)

0.120 0.269 0.070 0.067 0.035 0.020 Adj. R2


The t statistics are reported in the parentheses. a. Significance at 1%; b. Significance at 5%; c. Significance at 10%.

23

Table 5: Religious beliefs and constitutional commitment to social security (the dependent variable- an index of constitutional commitment to social security)

(6)

(5) (4) (3) (2) (1)

Constitutional commitment to:

Summary Index of Social Security

Work Injury

Old Age, Disability and

Survivors Sickness UnemploymentIncome Support

0.736 0.403 0.889 -0.019 0.755 1.487 Constant

(0.770) (0.460) (0.720) (-0.020) (0.580) (1.080)

-0.031 -0.045 -0.031 -0.006 -0.023 -0.062 Log GDP Per-Capita

(-0.340) (-0.550) (-0.270) (-0.070) (-0.190) (-0.480)

0.534 0.646b 0.005 0.904a 0.493 0.043 Catholic

(1.551) (2.060) (1.230) (2.563) (1.061) (0.087)

-0.001 0.111 0.002 0.325 -0.169 -0.987c Muslim

(-0.003) (0.302) (0.550) (0.788) (-0.311) (-1.708)

-0.291 -0.049 -0.002 0.102 -0.527 -0.750 Other

(-0.763) (-0.140) (-0.520) (0.261) (-1.022) (-1.369)

0.110 0.113 0.039 0.134 0.080 0.070 Adj.R²

64 64 64 64 64 64 Number of observations


24

Table 6: Religious beliefs and constitutional commitment to social security With Latin American dummy

(the dependent variable- an index of constitutional commitment to social security)

(6)

(5) (4) (3) (2) (1)

Constitutional commitment to

Summary Index of Social Security

Work Injury

Old Age, Disability and

Survivors Sickness UnemploymentIncome Support

0.368 -0.301 0.449 -0.231 0.375 1.521 Constant

(0.380) (-0.380) (0.360) (-0.230) (0.290) (1.070)

0.009 0.032 0.017 0.017 0.018 -0.066 Log GDP Per-Capita

(0.100) (0.420) (0.140) (0.180) (0.140) (-0.490)

0.003 0.002 0.003 0.008b 0.003 0.001 Catholic

(0.910) (0.850) (0.650) (2.100) (0.570) (0.120)

0.0003 0.001 0.003 0.003 -0.001 -0.009c Muslim

(0.090) (-0.550) (0.640) (0.840) (-0.250) (-1.700)

-0.003 0.0001 -0.002 0.001 -0.005 -0.007 Other

(-0.690) (0.04) (-0.450) (0.310) (-0.960) (-1.360)

0.481c 0.922ª 0.572c 0.276 0.496 -0.043 Latin America

(1.790) (4.170) (1.640) (0.980) (1.350) (-0.110)

0.142 0.306 0.066 0.133 0.095 0.056 Adj.R²

64 64 64 64 64 64 Number of observations


25

Table 7: A simple matrix of social security laws and constitutional rights Panel A: Old Age, Disability and Survivors (number of countries)

Social Security Law Constitutional Right Yes No Total

Yes 43 0 43 No 25 0 25 Total 68 0 68 Panel B: Unemployment (number of countries)


Yes 14 6* 20 No 33 15 48 Total 47 21 68 * The six countries are Bolivia, Dominican Republic, Panama, Nicaragua, India and Bahrain. Panel C: Work Injury (number of countries)


Yes 11 0 11 No 57 0 57 Total 68 0 68 Panel D: Sickness (number of countries)


Yes 13 2* 15 No 42 11 53 Total 55 13 68 * The two countries are Syria and Bahrain.

26

Table 8: Generosity of social security laws and constitutional commitment to social security

(4) (3) (2) (1)

Dependent variable:

Social Security Laws Index

(Botero et al, 2004)

Contribution Rate

(Old Age, Disability and

Survivors)

Contribution Rate(All programs)

Transfer Payment (% GDP)

-0.312 40.790ª 56.870ª -2.423 Constant

(-1.190) (3.090) (3.140) (-0.340)

0.089ª -4.061b -5.733ª -0.284 Log GDP Per-Capita

(2.650) (-2.370) (-2.430) (-0.310)

0.011c 1.327ª 2.159ª 1.592ª Population 65+

(1.870) (3.880) (4.590) (8.780)

0.034 2.145 2.750ª CCSS Summary Index

(1.330) (1.000) (3.330)

2.228c CCSS Old age, Disability and Survivors

(1.770)

0.489 0.204 0.261 0.758 Adj.R²

52 66 66 65 Number of Observations


27

Table 9: Generosity of social security laws and constitutional commitment to social security

(6) (5) (4) (3) (2) (1)

The dependent variable: an index of the extent of social security laws (Botero et al, 2004)

Social Security Laws

Index

Old Age, Disability and Death



Sickness and Health Unemployment

-0.656ª -0.171 0.173 -0.193 -0.178 0.639ª Constant

(-3.420) (-1.010) (-1.020) (-1.130) (-0.570) (2.840)

0.139ª 0.086ª 0.086ª 0.087ª 0.095ª 0.014 Log GDP Per-Capita

(6.680) (4.710) (4.690) (4.730) (2.820) (0.590)

0.005 CCSS Unemployment

(0.320)

0.056 CCSS Sickness

(1.360)

0.005 CCSS Old Age

(0.350)

-0.007

CCSS Old Age, Survivors, Disability

(-0.390)

0.037 -0.014 CCSS Summary Index

(1.390) (-0.610)

0.463 0.289 0.286 0.285 0.121 -0.043 Adj. R2



28

Table 10: Generosity of unemployment law and constitutional commitment to unemployment insurance

(5) (4) (3) (2) (1)

Dependent variable: an index of the extent of unemployment law (Botero et al, 2004)

Months of contributions or

employment required to qualify for

unemployment by law

Percentage of the worker’s

monthly salary deducted by law

to cover unemployment

Waiting period for

unemployment

Percentage of the net salary covered

by the net unemployment in case of a one-year

unemployment period

Summary index of

unemployment law

0.359 1.575ª 0.331 -0.255 0.639ª Constant

(0.990) (3.520) (0.610) (-0.420) (2.840)

0.058 -0.079c 0.059 0.071 0.014 Log GDP Per-Capita

(1.540) (-1.690) (1.040) (1.120) (0.590)

-0.038 0.011 -0.016 0.074c 0.005 CCSS Unemployment

(1.520) (0.380) (-0.420) (1.880) (0.320)

0.077 0.028 -0.017 0.064 -0.043 Adj. R2

39 39 39 36 39 Number of Observations


29

Table 11: Generosity of social security laws and transfer payments expenditures (Dependent variable - the share of transfer payments in GDP)

) 6( )5( )4( )3( )2( )1(

-2.925b-3.030b-2.145-1.609-0.6813.114Constant

(-2.350)(-2.170)(-0.880))-0.540()-0.130() 0.460(

1.588ª1.551ª1.680ª1.643ªª1.6351.628ªPopulation 65+

(9.820)(11.260)(9.280))10.880()10.860() 7.960(

-8.889UnemploymentBenefits

)-0.990(

-4.308Sickness andHealth Benefits

)-0.680(

-3.808Old Age, Disability and Death Benefits

)-0.730(

-3.377Social Security Laws Index

(-0.700)

0.023Sickness and maternity

(0.010)

-0.574Unemployment

(-0.330)

0.7190.7180.7270.7270.7180.622Adj ²R

656551514739ber ofNumObservations

The t statistics are reported in the parentheses a. Significance at 1%; b. Significance at 5%; c. Significance at 10%.

30

References Ben Bassat Avi and Momi Dahan, “Social and Economic Rights in the Constitution and in Practice,” August 2003, Hebrew University, School of Public Policy working paper #01.04 http://papers.ssrn.com/sol3/papers.cfm?abstract_id=407260 Botero Juan, Djankoy Dimeon, La-Porta Rafael, Lopez-de-Silanes Florencio and Shleifer Andrei, “The regulation of Labor,” Quarterly Journal of Economics, November 2004; 119(4): 1339-82. Finer, Samuel. The History of Government. Vol. I-III. Cambridge University Press, 1997. Glaeser Edward, Andrei Shleifer, Robert W. Vishny, “Legal Origins,” Quarterly Journal of Economics. 117(4), November 2002: 1193-1230. ILO, World Labour Report, Geneva 2000, Statistical Annex, Table 14. La-Porta, Rafael, Florencio Lopez-deSilanes, Andrei Shleifer and Robert W. Vishny, “The Quality of Government,” Journal of Law, Economics and Organization. April 1999; 15(1): 222-79. La-Porta, Rafael, Florencio Lopez-deSilanes, Andrei Shleifer and Robert W. Vishny, “Law and Finance,” Journal of Political Economy. December 1998; 106(6): 1113-55. La-Porta, Rafael, Florencio Lopez-deSilanes, Andrei Shleifer and Robert W. Vishny. “Legal Determinants of External Finance,” Journal of Finance. July 1997; 52(3): 1131-50. Mulligan, Casey B, Gil, Ricard and Sala-i-Martin, Xavier. “Social Security and Democracy.” NBER Working Paper Series. Cambridge, MA: May 2002. Mulligan, Casey and Sala-i-Martin, Xavier. “Internationally Common Features of Public Old-Age Pensions, and Their Implications for Models of the Public Sector.” Advances in Economic Analysis &Policy. Vol 4(1) 2004. Persson, Torsten and Tabellini, Guido (2003), The Economic Effect of Constitutions. MIT Press, Cambridge, MA. Persson, Torsten and Tabellini, Guido, (2004) “Constitutional Rules and Fiscal Policy Outcomes,” American Economic Review. March 2004; 94(1): 25-45. Reynolds, Thomas H., and Flores, Arturo A. Foreign Law: Current Sources of Codes and Basic Legislation in Jurisdictions of the World. Littleton, Colorado: Rothman 1989.

31

Rodrik, Dani. “Why Do More Open Economies Have Bigger Governments?” Journal of Political Economy. October 1998; 106(5): 997-1032. Social Security Throughout the World 2004 Tavares, Jose and Wacziarg, Romain. “How Democracy Affects Growth.” European Economic Review 45(2001) 1341-1378. World Bank Development Indicators, various years.

Documents

Constitutional Commitment to Social Security and Welfare Policy